Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Eq. (1.3.10) — the Miyazawa (rate) conservation principle

Proved
PalmQueueing.Palm.miyazawa_conservation

by mikedeng1 · Oct 1, 2026 · Mathlib 0df444a (Lean v4.33.1)

palm-calculuspoint-processqueueingrate-conservationstationary-process

The Miyazawa conservation principle. Let {Y(t)}\{Y(t)\}{Y(t)} be a bounded real-valued stochastic process, right-continuous with left-hand limits, and compatible with the flow {θt}\{\theta_t\}{θt​}. Let NNN be a point process compatible with {θt}\{\theta_t\}{θt​} and with non-null finite intensity λ\lambdaλ, and let {Y′(t)}\{Y'(t)\}{Y′(t)} be a real-valued stochastic process compatible with {θt}\{\theta_t\}{θt​} and such that

Y(1)=Y(0)+∫01Y′(s) ds+∫(0,1](Y(s)−Y(s−)) N(ds)(1.3.9)Y(1) = Y(0) + \int_0^1 Y'(s)\,ds + \int_{(0,1]} \big(Y(s) - Y(s-)\big)\, N(ds) \tag{1.3.9}Y(1)=Y(0)+∫01​Y′(s)ds+∫(0,1]​(Y(s)−Y(s−))N(ds)(1.3.9)

(for instance NNN counts all the discontinuity points of {Y(t)}\{Y(t)\}{Y(t)} and Y′(t)Y'(t)Y′(t) is the derivative of Y(t)Y(t)Y(t) between discontinuity points). Taking expectations on both sides and using the θt\theta_tθt​-invariance of PPP gives Miyazawa's formula

E[Y′(0)]+λEN0[Y(0)−Y(0−)]=0.(1.3.10)E[Y'(0)] + \lambda E^0_N[Y(0) - Y(0-)] = 0 . \tag{1.3.10}E[Y′(0)]+λEN0​[Y(0)−Y(0−)]=0.(1.3.10)

The mean drift of the continuous part is exactly cancelled by the mean jump rate: this is the (rate) conservation principle, applied many times in Chapter 3.

Boundedness of {Y(t)}\{Y(t)\}{Y(t)} is the hypothesis the book states. Remark 1.3.4 on p.24 observes that it can be replaced by any one of three integrability conditions; those are recorded in the mission's moderation notes rather than stated here, so that the item is the book's theorem and not a strengthening of it.

Formalization Note. Formula (1.3.10) treats E[Y′(0)]E[Y'(0)]E[Y′(0)] as a finite number, so the item assumes Y′(0)∈L1(P)Y'(0) \in L^1(P)Y′(0)∈L1(P). Lean's Bochner integral of a non-integrable function is 000, and without this hypothesis the statement would be false: with Y(t)=sin⁡(Ut+V)+(t−T0(t))Y(t) = \sin(Ut + V) + (t - T_0(t))Y(t)=sin(Ut+V)+(t−T0​(t)) over a lattice point process of random phase and UUU heavy-tailed (E∣U∣=∞E|U| = \inftyE∣U∣=∞), (1.3.9) holds pathwise while the Lean value of E[Y′(0)]E[Y'(0)]E[Y′(0)], for Y′(0)=Ucos⁡V+1Y'(0) = U\cos V + 1Y′(0)=UcosV+1, is 000 and λEN0[Y(0)−Y(0−)]=−1\lambda E^0_N[Y(0) - Y(0-)] = -1λEN0​[Y(0)−Y(0−)]=−1.

Preamble
import Mathlib
import Definitions.Def_PalmQueueing_Palm_PointProcess
Formal statement
/-!
# Eq. (1.3.10): the Miyazawa (rate) conservation principle (§1.3.3, pp.23-24)
-/

namespace PalmQueueing.Palm

open MeasureTheory

variable {Ω : Type*} [MeasurableSpace Ω]

/-- **The Miyazawa conservation principle**, Eq. (1.3.10) (§1.3.3, p.24). Let `{Y(t)}` be a
bounded real-valued corlol process compatible with the flow `{θ_t}`, let `N` be a point process
compatible with `{θ_t}` with non-null finite intensity `λ`, and let `{Y'(t)}` be a real-valued
process compatible with `{θ_t}` such that

`(1.3.9)  Y(1) = Y(0) + ∫_0^1 Y'(s) ds + ∫_{(0,1]} (Y(s) - Y(s-)) N(ds)`

— for instance `N` counts the discontinuity points of `{Y(t)}` and `Y'(t)` is the derivative of
`Y(t)` between them. Then

`(1.3.10)  E[Y'(0)] + λ E⁰_N[ Y(0) - Y(0-) ] = 0`:

the mean drift of the continuous part is exactly cancelled by the mean jump rate. This is the
*(rate) conservation principle*, which Chapter 3 applies repeatedly.

The boundedness of `{Y(t)}` is the hypothesis the book states; Remark 1.3.4 on p.24 records three
alternatives to it, recorded in `MODERATION_NOTES.md` rather than stated here. `hY'int` makes
explicit that `E[Y'(0)]` in (1.3.10) is a finite expectation, as the book's formula presupposes:
without it a Bochner integral of a non-integrable `Y'(0)` would silently read as `0`. -/
theorem miyazawa_conservation (S : PalmSetting Ω) (Y Yl Y' : ℝ → Ω → ℝ)
    (hYbdd : ∃ C : ℝ, ∀ (t : ℝ) (ω : Ω), |Y t ω| ≤ C)
    (hYmeas : ∀ t, Measurable (Y t)) (hYlmeas : ∀ t, Measurable (Yl t))
    (hY'meas : ∀ t, Measurable (Y' t)) (hY'int : Integrable (Y' 0) S.P)
    (hYright : ∀ (s : ℝ) (ω : Ω), ContinuousWithinAt (fun u => Y u ω) (Set.Ici s) s)
    (hYleft : IsLeftLimitProcess Y Yl)
    (hYcomp : IsCompatible S.θ Y) (hY'comp : IsCompatible S.θ Y')
    (hjump : ∀ ω, Y 1 ω = Y 0 ω + (∫ s in Set.Ioc (0 : ℝ) 1, Y' s ω ∂(volume : Measure ℝ))
        + ∫ s in Set.Ioc (0 : ℝ) 1, (Y s ω - Yl s ω) ∂(S.N.count ω)) :
    (∫ ω, Y' 0 ω ∂S.P) + S.lam * ∫ ω, (Y 0 ω - Yl 0 ω) ∂S.P0 = 0 := by sorry

end PalmQueueing.Palm
Source
Baccelli & Bremaud, Elements of Queueing Theory: Palm Martingale Calculus and Stochastic Recurrences, 2nd ed., Springer 2003, pp. 23-24, §1.3.3, Eqs. (1.3.9)-(1.3.10)
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me